Theory Of Artificial Neural Network And Its Application Research In Sediment Science

Computing the maximal grain size during incipient motion d_cmax and gross bed-load transport rate in Non-uniform sediment with a wide distribution are two important indexes for description its transport law, and it is also a core question for discussion in non-uniform sediment research. Artificial neural network (ANN) is a approximate simulation of biologic nerve system, which is a network model with a special algorithm got from biologic prototype after abstractly research. ANN is a adaptive and non-linear dynamic system made from superposition of many simple components which we call neuron.The question for computation of d_cmax belongs to incipient motion of sediment in essence, which is the same question with two different sides, because computation of critical condition (including critical velocity critical shear stress critical power) for a grain size under a bed-load condition is equal to computation of a grain size during incipient motion under the bed-load condition and the water flow condition given before. And computation of gross bed-load transport rate has great relationship with standard of incipient motion. In this dissertation, on the basis of comprehensive review of the study achievements on sediment transport, we get thefact that most of formula in sediment incipient motion and bed-load transport rate are established by classical regression analysis model, i.e. firstly establishing the formula according to related sediment transport theory, secondly computation for regression coefficients in formula by using field or laboratory experiment data. However, according to ANN's theory, making use of Visual Basic development tool, friendly-interface single output three layers' artificial neural network generator base on improved BP algorithm has been developed by the author, and after constructing the model, the value of d_cmax, Finally the gross bed-load transport rate of Non-uniform sediment with a wide distribution in flume experiment of stead sediment transportation have been forecasted.The main contents are as follows:Section-of theory of sediment transportation-on the basis of comprehensivereview of the study achievements on sediment transport, pros and cons in using classical regression model based on LMS during establishing formula in sediment incipient motion and bed-load transport rate; experiment data in Non-uniform sediment with a wide distribution in flume experiment of stead sediment transportation are been collected and coordinated.Section of ANN theory and its application study:After comparing study various model's algorithm, an improved BP algorithm has been adopted, i.e. a momentum factor a has been inducted. It can utilize the learning result of previous step during the training process, which can accelerate the training process of traditional BP algorithm.Realization of improved BP algorithm-Single output three layers' artificialneural network generator base on improved BP algorithm has been developed by the author, and the generator has some functions that the number of neuron in first and second layer and theirs related training parameters such as learning rate η. momentum factor a and the value of sum error e can all be self-defined by the users; Connection weights and threshold in each layer's neuron training data and teaching signals can also be input or modified in the friendly interface.The application study of improved BP algorithm in sediment science-usingtwo set of experiment data mentioned above, the value of d_cmax and gross bed-load transport rate in Non-uniform sediment with a wide distribution in flume experiment of stead sediment transportation have been forecasted by ANN generator. The data forecasted by network experiment data and data computed by classical regression model are compared and analysed from the comprehensive angel of sediment incipient motion theory and ANN theory.The conclusion of this dissertation-the value of d_cmax simulating bygenerator in Non-uniform sediment with a wide distr...